Endoscopic retrograde cholangiopancreatography (ERCP) in patients with altered surgical anatomy remains a big challenge for therapeutic endoscopist given lack of dedicated endoscope and accessories. There are several challenges encountered by the gastroenterologists when performing ERCP in patients with surgically altered anatomy such as identification of afferent limb, accessing and visualization of the papilla, and selective cannulation of the biliary and pancreatic ducts despite altered orientation of the papilla for therapeutic maneuvers. The challenge to reach the papilla could be due to multitude of factors such as the sharply angulated afferent limb entrance, endoscope looping, intra-abdominal postoperative adhesions, and long roux or afferent enteral limbs.^1–4 To overcome these difficulties, various strategies have been demonstrated to achieve high success rate in these technically challenging ERCPs. In this review we aim to discuss the most commonly encountered surgically altered anatomies requiring non-standard ERCP techniques.

Presently, Roux-en-Y gastric bypass (RYGB) is one of the most commonly encountered altered surgically anatomy with the least success rates among all post-surgical patients.⁵ The primary challenge is reaching the papilla and several approached have been studied and described to successfully accomplish the ERCP task in this patient population.

Enteroscopy assisted ERCP

Most commonly, in RYGB patients, the ERCP is attempted orally with the use of long length balloon assisted enteroscopes. Single balloon enteroscope (SBE), double balloon enteroscope (DBE), and rotational/spiral enteroscopes have all been studied with overall success rate of 70% in a large meta-analysis comprising 266 RYGB patients.⁵ There were no statistically significant differences between the therapeutic success rates between SBE, DBE and rotational enteroscope assisted ERCPs. Furthermore, there were no significant difference in the success rates between native papilla and anastomosis.⁵ Similarly, in a large U.S. multicenter retrospective study involving 63 RYGB patients, the overall success rate of ERCP was 62% with no significant difference among various enteroscopes. In this study, no difference in success rates was seen between RYGB and non-gastric bypass post-surgical anatomy patients.⁶ No difference in the mean procedural times were noted between SBE and DBE assisted ERCP either. The success rate of enteroscopy assisted ERCP does significantly drop if the roux enteral limb length is above 150 cm (25% compared to 88% in patient with shorter roux limb).⁷

It is important to note the availability of limited ERCP devices compatible with enteroscopes of 200-cm length. The short DBE with a working length of 152 cm allows the use of standard ERCP accessories and thus may prove to be extremely useful.^8,9 However, there are no studies comparing short and long DBE ERCP.

The overall major adverse event rate in the meta-analysis conducted by Skinner et al⁵ involving a total of 945 enteroscopy assisted ERCPs in post surgically altered patients including RYGB was 3.4% (n = 32). These included cholangitis (n = 1), pancreatitis (n = 11), bleeding (n = 3), perforation (n = 13), and death (n = 1). Six of the perforations required subsequent surgery. In the large U.S. multicenter study in patients with altered surgically anatomy who underwent an enteroscopy assisted ERCP, the overall major adverse event rate was 12.4% (n = 16/129). These included pancreatitis (n = 5), mild bleeding (n = 1), abdominal pain requiring hospital admission (n = 3), and throat pain requiring physician contact (n = 4), perforations (n = 2), and death (n = 1).⁶

Surgically assisted trans-gastric ERCP

If enteroscopy assisted ERCP has failed or in patients with roux limb length of more than 150 cm, transgastric ERCP can be performed using a hybrid surgical-endoscopic approach. Surgeon creates a gastrostomy into the excluded stomach which is then used to pass a standard duodenoscope to perform the ERCP either during same setting or 4–6 weeks later as a staged approach in non-urgent situations.^10,11 Even though this procedure is invasive, requires coordination between the surgeon and gastroenterologist and the patient’s supine position may pose some challenges, it may be useful in certain situations.¹ These include when patient needs an endoscopic ultrasound (EUS) assessment or cholecystectomy both of which can be performed during the same setting.¹ Furthermore, this procedure has been shown to cost beneficial in-patient with long roux limb > 150 cm.⁷

In order to circumvent the inconvenience of coordinating with surgeons and performing procedure in un-familiar territory, Baron¹² reported DBE assisted retrograde placement of a percutaneous endoscopic gastrostomy (PEG) to allow subsequent ERCP in a patient with a long-limb RYGB. Thompson et al¹³ have further demonstrated a single session ERCP approach by EUS-guided percutaneous gastric access. This approach involved accessing the gastric remnant using a 22 G needle under EUS guidance followed by insufflation of contrast and carbon dioxide to distend the wall of gastric pouch. A fully covered esophageal stent or PEG is then placed percutaneously after affixing the gastric pouch to the anterior abdomen wall using T-fasteners. ERCP can be performed by passing a duodenoscope through the esophageal stent.

Success rates of the hybrid procedure in several case series have been reported to be as high as 90%–100%. Saleem et al¹⁰ reported achievement of cannulation and therapeutic interventions including biliary and pancreatic sphincterotomy in all 15 cases with laparoscopic assisted ERCP in RYGB anatomy without any postoperative complications. In a large retrospective review comparing data on balloon enteroscopy-assisted ERCP (n = 32) versus laparoscopy-assisted ERCP (n = 24) in RYGB patients, hybrid procedure was superior in papilla identification (100% vs 72%), cannulation rate (100% vs 59%), and therapeutic success (100% vs 59%). The total procedure time was shorter but the endoscopist’s time was longer in enteroscopy assisted ERCP. There was no significant difference in post-procedure hospital stay or complication rate between the two groups.⁷ Majority of the adverse events with this procedure are related to the gastrostomy tube itself rather than the ERCP.^7,14

Gastro gastric fistula assisted ERCP

With continued advancement in the therapeutic role of EUS, several innovative EUS assisted ERCP techniques have been described in RYGB patients. The most promising among those is the use of recently approved fully covered lumen apposing metal stent to create an internal gastro-gastric fistula which then allow passage of the standard duodenoscope to perform the ERCP. The fistula is then closed using endoscopic suturing after performing the ERCP.^15,16 This may be a preferred if the RYGB patient has a pre-existing gastro-gastric fistula. A duodenoscope could be passed through the fistula after either balloon dilatation or placement of fully covered enteral esophageal stent. The major concerns with this procedure are the risk of weight regain especially in case of failure to successfully close the gastro-gastric fistula completely and the risk of stent migration.

Similar to RYGB, the access to hepatico-jejunostomy anastomosis in post Roux-en-Y hepaticojejunostomy (RYHJ) requires traversing the roux and afferent enteral limbs. However, unlike the RYGB, patients with this surgery have intact duodenum, major papilla and native pancreatic anatomy. Hence a non-standardized ERCP approach is required for biliary access alone. It requires either a colonoscope or balloon assisted enteroscope to perform ERCP for biliary indication (Fig. 1). The efficacy and adverse event rate was seen to be similar to that observed in RYGB patients.^5,6

In both classic and pylorus preserving Whipple surgery (pancreato-duodenectomy), the length of the afferent limb containing the choledocho-jejunal anastomosis is short spanning 40–60 cm. The pancreato-jejunal anastomosis is usually 10 cm distal to the biliary anastomosis. Hence the anastomoses can be reached using a duodenoscope, gastroscope or a colonoscope (Fig. 2).^1–3 In a single tertiary care center study involving 51 Whipple resection patients using a duodenoscope in the majority, an overall 51% technical success rate was demonstrated. Success was significantly higher for biliary indications (84%) than for pancreatic indications (8%). When both ductal systems were aimed at, success was 72%. The mean duration of procedure was 64 minutes and complications (retroperitoneal perforation n = 1 and Mallory Weiss tear n = 1) occurred in 2% of the procedures.¹⁷

Enteroscopy assisted ERCP has also been studied and described in post-Whipple surgery patients. In a systemic review including patients from 18 studies, the success rate of ERCP in this patient population using enteroscopy was 76%.⁵ Shah et al⁶ reported a 64% ERCP success rate in a multicenter study involving 31 post-Whipple resection patients when evaluating the efficacy of enteroscopy assisted ERCP. The cumulative adverse event rate with enteroscopy assisted ERCP has already been discussed before.

In Billroth type II gastrectomy, the distance between the gastro-jejunal anastomosis and major papilla is short (approximately 30–50 cm) similar to post-Whipple patients thus allowing the use of duodenoscope, gastroscope or a colonoscope to perform the ERCP. The primary challenge lies in the cannulation due to the altered orientation of biliary orifice (usually caudal). Braun anastomosis, a jejunojejunostomy between the afferent and the efferent limbs, accompanying Billroth II operation may add additional challenges in terms of identifying the enteral limb, longer afferent limb and sharp angulations.^1,4

The success rate of ERCP in Billroth II anatomy is in the range of 80%–90%.^5,18–20 In a large retrospective study involving 713 patients with Billroth II reconstruction, Bove et al¹⁸ reported successful duodenal intubation rate of 86.7% with a successful cannulation rate of 93.8% among intubated patients. The adverse event rate was 4.3% with an overall mortality rate of 0.3%. In a prospective head on randomized controlled trial (n = 45), cannulation rate was significantly higher using duodenoscopes (87%) as compared to gastroscopes (68%). However, the perforation rate was significantly lower with the forward viewing endoscopes.²¹

In order to improve orientation for cannulation in Biliroth II, several ERCP accessories such as a rotatable sphincterotome, a sigmoid shaped sphincterotome or a straight catheter with groomed distal tip have shown good efficacy (~90%) with low adverse event rates (< 5%).^22,23 Another popular strategy is the use biliary stent as a guide to perform needle knife sphincterotomy with a reported success rate of 83% and early complication rate of 39% (bleeding 16%, pancreatitis 8%, perforation 2%) in a small cohort of 18 patients.²⁴ In above study, a randomized comparison of trans-papillary balloon dilatation (n = 16) with endoscopic sphincterotomy (n = 18) in Billroth II patients showed similar overall success rates for stone removal with balloon dilation (88%) but without any post-procedure bleeding events. Endoscopic papillary large balloon dilation after minor sphincterotomy was also shown to be an effective (100%) and safe (no bleeding, pancreatitis, and perforation) method for removal of bile duct stones in a study involving 26 patients with bile duct stones and history of Billroth II gastrectomy.²⁵

EUS-guided biliary drainage (EUS-BD) has been extensively reported after first description in 2001 by Giovannini et al.²⁶ In a meta-analysis involving 74 reported cases of EUS-BD in surgically altered anatomy, Siripun et al²⁷ reported a pooled technical success, clinical success, and complication rates of 89.2%, 91.1%, and 17.5%, respectively. More recently, an international multi-center cohort study compared enteroscopy-assisted ERCP vs EUS-BD in patients (n = 98) with surgically altered anatomy. EUS-BD group had higher technical (98% vs 65.3%) and clinical success (88% vs 59.1%) rates as compared enteroscopy assisted ERCP group. Procedural time was significantly shorter in the EUS-BD group (55 minutes vs 95 minutes). However, adverse events occurred more commonly (20% vs 4%) and the length of stay (6.6 days vs 2.4 days) was longer in the EUS-BD group.²⁸

ERCP in a surgically altered anatomy is a technically challenging undertaking. Several strategies to improve the success rate have been recommended by various endoscopy experts. It starts with adequate pre-procedural planning by reviewing the surgical notes and imaging in order to understand the type of surgery and length of enteral limbs needed to be traversed. This should allow the choice of an appropriate approach (such as trans-oral vs trans-gastric) and type of endoscope (depending upon the length and angle of view) needed to perform the procedure. It is important to keep in mind the compatibility of ERCP accessories when using non-standard endoscopes.

Fluoroscopic guidance (endoscope in right upper quadrant), amount of bile (greater in the afferent limb) and direction of peristalsis can all be utilized in order to identify the pancreato-biliary afferent limb. Repositioning the patient to either supine or left lateral and/or application of abdomen pressure may be required in order to prevent endoscope looping and aid in its advancement. When using forward viewing endoscopes, use of clear cap has been advised especially in patients with altered pancreato-biliary anastomosis.

After a successful or failed ERCP attempt, it may be helpful to tattoo the afferent limb in case the patient needs a repeat procedure. Furthermore, it is also important to document a detailed procedure note with accurate measurements and description of the technique.^1–4